Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Building Environment and Energy Engineering | en_US |
| dc.contributor.advisor | Xiao, Fu Linda (BEEE) | en_US |
| dc.creator | Zhang, Hanbei | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/14175 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Development of stochastic occupancy modelling methods and occupancy-integrated mpc for smart built environment control and its implementation | en_US |
| dcterms.abstract | Modern HVAC systems, which dominate building energy use, face the critical challenge of balancing the conflicting objectives of thermal comfort, indoor air quality (IAQ), and stringent energy efficiency targets. Among the various factors, occupant behavior has emerged as one of the most significant drivers of building energy use. However, traditional building automation systems often rely on oversimplified assumptions about occupancy patterns, resulting in substantial energy wastage due to oversupplying unoccupied spaces, while also risking compromised occupant comfort and satisfaction. | en_US |
| dcterms.abstract | With the advancements of Internet of Things (IoT) devices and information communication technologies in smart buildings, real-time occupancy information—such as presence, number of occupants, and personalized feedback—is increasingly accessible. Additionally, IoT-enabled built environment data is being extensively collected, providing unprecedented opportunities for data-driven building management. However, there remains a significant gap in the development of data-driven modeling and occupancy-integrated control systems capable of effectively leveraging IoT data to match building energy services with actual occupancy demand. Traditional data-driven modeling methods often suffer from high model complexity and poor generalization ability, limiting their scalability and applicability across diverse building spaces. At the same time, the suboptimal performance of simple, reactive control strategies highlights the need for the advanced stochastic occupancy-integrated model predictive control (MPC) strategy. Furthermore, there is a lack of real-world implementation and validation of MPC systems under living conditions with active occupant participation and feedback. | en_US |
| dcterms.abstract | This research aims to address the aforementioned challenges by developing advanced stochastic occupancy modeling methods to accurately identify diverse occupancy patterns, integrating stochastic occupancy prediction into the MPC strategy to align building energy system control with real occupancy demand, and implementing the MPC strategy in an IoT-enabled real-world living lab environment to evaluate its practical impact. | en_US |
| dcterms.abstract | This study first proposes the Adaptive B-Spline-based inhomogeneous Markov chains (IMC) method for stochastic occupancy modelling. This method introduces a dynamic knot adjustment mechanism to better capture occupancy variations across different types of spaces. The proposed method achieves significantly improved prediction accuracy while reducing model complexity. (Chapter 3) | en_US |
| dcterms.abstract | To address the poor generalization issue, this study further develops a band structure-integrated continuous-time inhomogeneous Markov chain (CTIMC) modelling method. The band structure constrains transitions between states to physically meaningful neighboring states, reflecting realistic occupant movement and behavior. Compared to traditional discrete-time IMC models, the band structure-integrated CTIMC method demonstrates superior generalization ability and computational efficiency. (Chapter 4) | en_US |
| dcterms.abstract | This research further integrates the stochastic occupancy prediction into the MPC strategy for multi-objective optimal built environment control. The stochasticity and time-inhomogeneity of occupancy heat gains and CO₂ generations are embedded in the prediction of built environment and energy consumption in MPC. TRNSYS simulation demonstrates the effectiveness of the stochastic occupancy-integrated MPC in achieving significant energy savings while improving thermal comfort and IAQ. (Chapter 5) | en_US |
| dcterms.abstract | This study further develops an IoT-enabled architecture for intelligent built environment management, integrating stochastic occupancy modeling and real-time data into an Occupant-in-the-loop MPC strategy and validates through field implementations. Deployed in a living lab, the system optimizes HVAC operations, balancing thermal comfort, air quality, and energy efficiency. A one-month experimental evaluation showed 54.9% energy savings, improved comfort, and sufficient ventilation, advancing real-world MPC implementation and future IoT-based building management enhancements. (Chapter 6) | en_US |
| dcterms.abstract | Finally, to enhance the scalability and cost-effectiveness of strategy deployment, this study develops an IoT retrofit scheme that considers the trade-off between costs and model accuracy. A data-driven modeling method using low-cost data is proposed, with systematic analysis conducted on model accuracy and the associated costs. The results indicate that the proposed modeling method based on low-cost data can achieve prediction accuracy comparable to modeling methods using high-cost, high-accuracy data. The proposed method facilitates large-scale, cost-effective deployment of the intelligent built environment management system. (Chapter 7) | en_US |
| dcterms.abstract | This thesis provides a comprehensive framework for intelligent built environment management, combining cutting-edge stochastic modeling, predictive optimization, and real-time occupant interaction. By addressing the challenges posed by dynamic and uncertain occupancy patterns, this research contributes to more sustainable, scalable, and occupant-centric building operations. The findings have broad implications for a range of building types and operational scenarios, offering new insights into the integration of data-driven modeling, IoT, and advanced control technologies to optimize the built environment. | en_US |
| dcterms.extent | xvi, 187 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2025 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Buildings -- Energy consumption -- Data processing | en_US |
| dcterms.LCSH | Demand-side management (Electric utilities) | en_US |
| dcterms.LCSH | Predictive control | en_US |
| dcterms.LCSH | Internet of things | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/14175